Page 49 - Kaleidoscope Academic Conference Proceedings 2024

Innovation and Digital Transformation for a Sustainable World

have a specific business meaning in a certain format. These each configuration detail, the system checks whether the
fields require rules to be set for converting into proportion details trigger any rules. If a rule is triggered, the system
values before subsequent comparisons. In Case (5), the automatically records key information for that configuration
oversubscription ratio is a field that satisfies a specific format. detail, retrieves the corresponding handling advice for the
In actual rule comparisons, this field needs to be converted inspection rule, and consolidates the relevant information.
into a floating-point number before further comparison. The final result is summarized as the overall outcome of an
Additionally, Formula (12) utilizes two extension functions, inspection. The system automatically persists the results in
Formula (2) and Formula (5), combining multiple the database for later full export.
mechanisms. This is also a feature that the dynamic
inspection method needs to support.
Scenario 6: business based comparison. This scenario is
closely related to CMDB business, involving the business
correspondence of detailed fields. The related business
relationships depend on specific business scopes. Therefore,
rules need to be set to satisfy specified business scope
associations, as in case (6). In this rule, the operating system
type is independently maintained in a business collection
table in CMDB, forming a business correspondence
relationship. For example, the SUSE operating system (a Figure 3 – CMDB data inspection process
Linux distribution maintained by the German company SuSE
Linux AG) may be reported to the CMDB system with Due to the constant changes in configuration data within
variations such as 'SUSELinux,' 'SUSE Linux,' 'suse,' 'SLES,' CMDB, it is necessary to generate snapshots of the entire
due to differences in the reporting source's underlying configuration items from CMDB and persist them before
management technology or variations in the configuration each inspection task starts. This facilitates later analysis by
content collection format. Similar variations may occur for personnel, who can use the inspection results in conjunction
other types of operating systems. To address this, CMDB with snapshots of configuration details at the time of
maintains a functionality called the business mapping list for inspection to further investigate data issues. This, in turn,
such scenarios. Formula (13) leverages the extensions encourages data providers to rectify problematic data. To
provided by formulas (6) and (7) to achieve the final rule alleviate pressure on the product environment, the system
semantics. incorporates the scheduled task framework. It performs
inspections on key resource types every day, displaying the
2.2 The application of dynamic inspection method in results on the interface and offering download options for
CMDB business both results and snapshots. To expedite the resolution of
issues identified during inspections, we provide additional
The CMDB in China Mobile's IT Cloud serves as a features such as the duration of the problem, trend charts, and
configuration management database, playing the role of a other supportive functions. These features further contribute
data source in this system. The inspection system starts to the assessment of data quality and propel the progress of
inspections by retrieving data from CMDB. Dynamic data rectification.
inspection rules are applied in CMDB with the aim of
establishing a mechanism that covers existing models, is 2.2.2 Application of the inspection method in CMDB
compatible with subsequent additions of models, and adapts
to dynamic changes in model fields and inspection rule (1) Data extraction for inspection. In essence, the input of
extensions. This is to accommodate the evolving resource the inspection system is the detailed configuration data
models and detailed field requirements, addressing issues existing in the CMDB. Due to the dynamically changing
such as diverse data sources leading to poor data quality and business requirements, the configuration item model
resolving quality problems stemming from outdated undergoes dynamic changes. Along with the dynamic
configuration item data in legacy management systems. changes in the model, the rules also change accordingly.
Therefore, to cope with the frequently changing
2.2.1 Overall design of inspection system requirements, the inspection system needs to extract CMDB
data through a low-code approach. Specifically, the content
We developed the system based on the Spring Boot we ultimately need to extract includes two parts: a) all fields
framework[13], adopting containerized management[14] involved in inspection rules under the required inspection
and deploying it in a distributed manner using model; b) fields not involved in inspection rules but needed
Kubernetes[15]. The overall logic is illustrated in Figure 3. for the final presentation of results.
The system interacts with CMDB through OpenAPI[16],
fetching data in batches for various models such as bare (2) Execution of dynamic inspection rules. After obtaining
metal, virtual servers, hosts, and various network devices. the source data, we need to retrieve the inspection rules for
Subsequently, the system retrieves corresponding inspection this batch of data in order to inspect these source data.
rules for each model, which are maintained in the database Furthermore, the inspection rule information we obtain
for easy access at any time. Applying the inspection rules to needs to meet the following conditions: a) Corresponding

– 5 –